Telomere length changes in patients with aplastic anaemia


: Hoon Kook, M.D., Department of Paediatrics, Chonnam National University Medical School, 8 Hak-Dong, Dong-Ku, Kwangju 501-757, South Korea. E-mail:


To investigate telomere changes in patients with aplastic anaemia (AA) and clinical factors influencing the telomere dynamics, telomere length (TL) was measured in peripheral blood mononuclear cells using Southern blot analysis of 42 patients with AA and 39 healthy normal controls. Nineteen patients received supportive treatment only, while the remaining 23 patients received immunosuppressive therapy with anti-thymocyte globulin or anti-lymphocyte globulin ± cyclosporin A. In AA patients, TL was on average 1·41 kb shorter than that of age-matched normal controls (P < 0·001). In patients treated with immunosuppression, the mean TL of non-responders was significantly shorter than that of age-matched normal controls (P < 0·001), while no difference in TL was detected in responders compared with controls. Positive correlation was observed between the extent of telomere shortening, the severity of neutropenia (P = 0·05) and the degree of mean corpuscular volume elevation (P = 0·005) at the time of the study. However, there was no correlation with time elapsed since diagnosis (P = 0·214). These findings suggest that haematopoietic stem cells in patients with AA rapidly lose TL at the onset of the disease. The TL shortening may reflect the severity of impairment of haematopoiesis.

Aplastic anaemia (AA) is characterized by a severely contracted haematopoietic stem cell pool in the bone marrow owing to an acquired intrinsic stem cell defect and/or an immune destruction of haematopoietic stem cells (Young & Maciejewski, 1997; Young, 1999). Increased loss of the primitive haematopoietic progenitor cells in AA is compensated (if not fully) by recruitment of more mature progenitor cells and their more frequent divisions (Ball et al, 1998). The extra demand imposed on the remaining stem cells may translate into more replicative stress compared with normal individuals, resulting in an accelerated telomere shortening in patients with AA (Ball et al, 1998; Wynn et al, 1998a).

Telomeres, consisting of specific proteins/tandem repeat DNA, act to stabilize the chromosome ends and infer protection against enzymatic end-degradation (Blackburn, 1991; Zakian, 1995). Telomere length (TL) is maintained by a balance between replication rate and telomerase activity (de Lange, 1998). Telomeric DNA decreases by 50–100 bp with each somatic cell division and progressively shortens with increasing age both in vitro and in vivo (Hastie et al, 1990; Harley, 1991; Blackburn, 1994; Rhyu, 1995; Moyzis et al, 1998). It has been postulated that TL in haematopoietic stem cells may serve as an indicator of haematopoietic ageing and a molecular measure of the remaining replicative potential (Harley, 1991; Allsopp et al, 1992; Slagboom et al, 1994; Vaziri et al, 1994; Engelhardt et al, 1997; Frenck et al, 1998; Wynn et al, 1998a).

Secondary late clonal diseases such as myelodysplastic syndrome (MDS), paroxysmal nocturnal haemoglobinuria (PNH) or leukaemia can develop in AA, in up to 50% of patients successfully treated with immunosuppression (Socie et al, 1993; Tichelli et al, 1994). Moreover, despite haematological recovery after immunosuppressive therapy (IST) or bone marrow transplantation, the quantitative as well as qualitative defects persist in both primitive haematopoietic stem cells and mature haematopoietic progenitors (Maciejewski et al, 1994, 1996; Scopes et al, 1994). Although the causes of these secondary clonal diseases have not been delineated, the persistent underlying pathophysiology of marrow failure, rapid stem-cell expansion, the failure of immune surveillance in malignant cells or the oligoclonality itself may all contribute to the evolution (Young, 1995). In addition, the accelerated telomere shortening, causing genetic instability as well as replicative senescence, may predispose patients with AA to the late clonal disorders (Allsopp et al, 1992; Shay et al, 1996; Ball et al, 1998; Wynn et al, 1998a). A few recent studies have demonstrated the progressive decrease of TL in peripheral blood granulocytes and mononuclear cells (MNCs) of AA patients (Ball et al, 1998; Brümmendorf et al, 1999). A similar finding of accelerated telomere shortening has been demonstrated during haematological reconstitution in haematopoietic stem cell transplantation recipients, suggesting the possibility of the earlier onset of late clonal haematopoietic disorders (Notaro et al, 1997; Shay, 1998; Wynn et al, 1998b; Lee et al, 1999; Wynn et al, 1999).

In this study, we measured TL in peripheral blood MNCs of 42 AA patients and attempted to correlate the extent of telomere shortening (ΔTEL) with several clinical characteristics of the patients.

Patients and methods

Patient characteristics Forty-two patients treated for idiopathic acquired AA at three different hospitals in Korea between April 1998 and February 1999 were included in this study (Table I). Their age ranged between 5 years and 79 years with a mean of 29·6 years. Among them, 19 patients received only supportive treatment, such as transfusions or androgens, while 23 patients received IST with anti-thymocyte globulin or anti-lymphocyte globulin ± cyclosporin A. Patients who had undergone allogeneic bone marrow transplantation and those with secondary clonal disorders, such as MDS or PNH, were excluded from the study. Patients with supportive treatments were classified into two groups, severe AA (SAA, n = 10) and non-severe AA (NSAA, n = 9), according to the severity of haematological findings at the time of the study. SAA was defined by the standard criteria: less than 25% marrow cellularity and depression in at least two out of three blood counts [corrected reticulocyte < 1%; absolute neutrophil count (ANC) < 0·5 × 109/l and platelets < 20 × 109/l] (Camitta et al, 1976). In addition, patients who did not fulfil the severe criteria but who had hypocellular bone marrow with low blood counts were considered as NSAA. We used the following functional response criteria after IST: a complete response (CR) was defined as an ANC of 1·0 × 109/l or greater, a platelet count of 100 × 109/l or greater and transfusion independence; a partial response (PR) was defined as a ANC of 0·5 × 109/l or greater, a platelet count of 30 × 109/l or greater and transfusion independence; patients who remained transfusion dependent were classified as non-responders, regardless of the neutrophil and platelet counts (Doney et al, 1997; Marsh et al, 1999). To compare mean TL with response to IST, we divided the treated patients into responders (CR or PR) (n = 7) and non-responders (n = 16). There were no differences in the severity or the duration of disease between the supportive treatment group and the immunosuppressive therapy group.

Table I.  Clinical characteristics and telomere changes of aplastic anaemia patients.

(× 109/l)
(× 109/l)
  1. ANC, absolute neutrophil count; Hb, haemoglobin; MCV, mean corpuscular volume; SPT, supportive treatment; IST, immunosuppressive therapy; SAA,severe aplastic anaemia; NSAA, non-severe aplastic anaemia.

 SAA1047·5 ± 7·70·74 ± 0·187·4 ± 0·6104·7 ± 3·523·7 ± 6·637·7 ± 21·41·20 ± 0·24
 NSAA941·7 ± 7·41·63 ± 0·309·3 ± 0·8107·9 ± 3·390·9 ± 28·431·9 ± 14·51·28 ± 0·15
 Non-responder1621·3 ± 3·10·60 ± 0·106·1 ± 0·4100·0 ± 2·917·4 ± 3·245·6 ± 9·41·68 ± 0·19
 Responder731·6 ± 9·82·16 ± 0·2511·1 ± 0·8102·6 ± 3·587·3 ± 16·823·7 ± 6·11·24 ± 0·26
Total4233·6 ± 3·51·11 ± 0·137·9 ± 0·4103·3 ± 1·747·0 ± 8·537·2 ± 6·91·40 ± 0·11

Normal controls Blood samples were collected from 39 normal individuals ranging from newborn to 72 years of age with normal blood counts.

TL measurement TL analyses were carried out as reported previously (Lee et al, 1999). In brief, all peripheral blood samples from patients and normal controls were collected after informed consent was obtained and stored at −80°C until DNA extraction. Peripheral blood MNCs were separated by centrifugation using Ficoll-Hypaque. High-molecular-weight DNA was extracted from 1 to 3 × 107 MNCs using 1 ml of DNAZOL reagent and the integrity of extracted DNA was confirmed using 1% agarose gel electrophoresis. RsaI-digested DNA (4 μg of each sample) was size-fractionated by electrophoresis on 0·6% agarose gels. DNA was transferred onto a nylon membrane (Hybond N+, Amersham, UK) using the vacuum transfer method for Southern blotting. The filters were hybridized with a biotinylated telomere probe (TTAGGG)4 (TeloQuant, PharMingen, San Diego, CA, USA) in hybridization buffer overnight at 65°C and the hybridized probe was detected using the chemiluminescence method. Telomere lengths were assessed quantitatively by densitometric analysis of autoradiographs using the transmitter scanning videodensitometer (Zenith Video Densitometer, Biomed, Fullerton, CA, USA). The mean TL in each sample was then identified as the peak intensity of the TL in kb using densitometry (Fig 1).

Figure 1.

A representative Southern blot analysis of telomere length using the (TTAGGG)4 probe hybridized against Rsa 1-digested genomic DNA from peripheral mononuclear cells of patients with aplastic anaemia (AA). Vertical axis shows telomere length (kb). Lanes 1–5 show DNA obtained from the following patients with AA: 1, 17 years; 2, 29 years; 3, 37 years; 4, 53 years, 5, 55 years.

Statistical analysis Values are presented as mean ± SEM. The Mann–Whitney U-test was used to compare the TL between subgroups. Linear regression analyses were performed to assess the correlation between ΔTEL and several clinical factors.


Mean TL in healthy normal controls

According to our previous measurements, the mean TL in peripheral blood MNCs from normal controls was 9·68 ± 0·33 kb (range, 5·65–14·40 kb) and there was a progressive shortening of TL with age (Lee et al, 1999). The TL of normal controls was plotted using the following equation: T = 10·86 − 0·0384 × A, in which T = TL in kb and A = age in years.

Telomere shortening in patients with AA

Compared with age-matched normal controls (mean, 9·56 ± 0·14 kb), the mean TL in patients with AA was decreased by 1·41 kb (mean, 8·15 ± 0·16 kb; range, 6·05–10·45) (P < 0·001) (Fig 2). The degree of telomere shortening in AA patients was equivalent to 36·7 years of ageing in the age-matched normal control, according to linear regression analysis. The mean ΔTEL was 1·24 kb in the supportive treatment group and 1·54 kb in the IST group (P = 0·133).

Figure 2.

Comparison of telomere length between age-matched normal controls and AA patients. Telomere lengths in AA (8·15 ± 0·16) were significantly shorter than those of normal controls (9·56 ± 0·14)(P < 0·001).

Among patients who received supportive treatment only, there was no difference in the mean ΔTEL between the SAA group (n = 10; 1·20 ± 0·24 kb) and NSAA group (n = 9; 1·28 ± 0·15 kb) (P = 0·462). Similarly (Table I), there was no significant difference in mean ΔTEL between non-responders (1·68 ± 0·19 kb) and responders (1·24 ± 0·26 kb) after IST (P = 0·385). When compared with controls, however, the mean TL of non-responders to IST (n = 16; 8·36 ± 0·26 kb) was significantly shorter than that of age-matched normal controls (10·04 ± 0·12 kb) (P < 0·001), whereas the TL in responders to IST (n = 7; 8·41 ± 0·53 kb) was not statistically different from that of age-matched normal controls (9·65 ± 0·38 kb) (P = 0·085) (Fig 3).

Figure 3.

Comparison of mean telomere length between the response groups for IST and age-matched normal control groups. NC = putative normal controls.

Influences of clinical factors on TL shortening

We tried to correlate the ΔTEL with several clinical and haematological factors relevant to AA. ΔTEL showed a negative correlation with ANC in peripheral blood at the time of the study (r = −0·304; P = 0·05; Fig 4) and showed a positive correlation with MCV levels (r = 0·429; P = 0·005; Fig 5). However, haemoglobin level (P = 0·240) and platelet count (P = 0·304) showed no significant correlation with ΔTEL. Moreover, duration of the disease was not correlated with ΔTEL either in all AA patients or in patients with supportive treatment only (P = 0·214).

Figure 4.

Correlation of the changes of telomere length (ΔTEL) with absolute neutrophil counts ( × 109/l) in AA patients (r = −0·304; P = 0·05).

Figure 5.

Correlation of the changes of telomere length (ΔTEL) with MCV levels (fl) in AA patients (r = 0·429; P = 0·005).


Telomeres are the ends of eukaryotic chromosomes composed of simple tandem hexameric repeats bound to specific proteins and are essential for the stability of chromosomes and genes (Blackburn, 1991, 1994; Zakian, 1995). In humans, telomeric DNA contains 5–15 kb of TTAGGG repeats at the end of all chromosomes (de Lange, 1998). As an end-replication problem during progressive cell division may result in telomeric loss, TL serves as an indicator of the replicative history of cells and may be the molecular counterpart of replicative potential remaining in cells (Harley, 1991; Allsopp et al, 1992; Shay et al, 1996). However, there is an inherent variation in telomere length between normal individuals and within different blood cell types of an individual, such as memory and naive T-cells, and granulocytes and lymphocytes (Iwama et al, 1998; Rufer et al, 1999; Zeichner et al, 1999).

Critically shortened telomeres may induce chromosome fusion or massive genomic instability, eventually contributing to age-related clonal disorders (Shay et al, 1996; Curtis et al, 1997; Wynn et al, 1998a). Although haematopoietic stem cells contain telomerase activity, the loss of telomeric sequences cannot be entirely prevented because the levels of telomerase are not sufficient to keep pace with the replicative loss (Weng et al, 1995; Iwama et al, 1998; Wynn et al, 1998a).

The long-term culture-initiating cells (LTC-ICs) in peripheral blood and bone marrow were profoundly decreased in patients with AA (Maciejewski et al, 1996). These findings suggest that the contracted stem cell pool in patients with SAA may be compensated by increased turnover of the stem cells, eventually leading to accelerated telomere shortening. Consistent with earlier observations (Shay et al, 1996; Ball et al, 1998), we found significant telomere shortening of MNCs in AA patients. The telomere shortening in AA might occur at the level of haematopoietic stem cells or result from the compensatory increase of more mature progenitor cells required to offset the loss of more immature cells (Ball et al, 1998).

Previously, Ball et al (1998) reported that the rate of telomere loss was apparently stabilized in AA patients with complete haematological recovery compared with those without recovery. In the present study, the TL of non-responders after IST was significantly decreased compared with that of normal controls. The TL of responders also tended to be shorter than that of normal controls, although the difference was not statistically significant, presumably owing to the small numbers in the group (N = 7). The finding of decreased TL despite full haematological recovery of the blood counts may be explained by the fact that the LTC-IC numbers of recovered AA after IST remained below those of normal controls (Maciejewski et al, 1996). This finding might also be associated with the propensity for development of late secondary clonal diseases such as PNH, MDS or leukaemia in patients after recovery from IST (Socie et al, 1993; Tichelli et al, 1994).

One of the goals of this study was to evaluate the correlation between clinical factors and changes in TL. The extent of telomere shortening in patients with AA showed a positive correlation with the degree of neutropenia and the mean corpuscular volume increase at the time of the study. These findings suggested that the degree of telomeric loss closely correlated with the impairment of haematopoiesis in AA. However, there was no correlation between ΔTEL and haemoglobin levels or between ΔTEL and platelet counts.

In the current study, there was no correlation between ΔTEL and the duration of the disease, suggesting that TL might shorten rapidly during the onset of AA and the rate of shortening after the initial insult might be similar to the rate in healthy subjects. However, Ball et al (1998) reported that telomere loss in active AA was correlated with the duration of the disease. Reasons for this discrepancy are not clear, but several potential factors including differences of disease duration, differences in number of patients receiving IST, individual heterogeneities of telomere length and limited numbers of study patients should be considered.

There were several limitations in interpreting telomere dynamics of bone marrow failure patients in the current study. First, only a limited number of AA patients, most of whom had a relatively short duration of the disease, were studied. Second, sequential measurements of TL in the same patients should be performed as there might be considerable telomeric heterogeneity between individuals. Third, the telomere length in different cell populations, such as granulocytes, may be studied (Rufer et al, 1999). Fourth, a direct connection of telomere shortening with clonal evolution could not be addressed in this study.

In conclusion, our findings suggest that haematopoietic stem cells in patients with AA rapidly lose telomere length at the onset of the disease. The telomere shortening may be further accelerated depending on the severity of stem cell pool contraction. Further studies of telomere loss in AA are needed to address the relationship with late clonal disorders.


The authors thank Dr J. P. Maciejewski, NHLB1, N1H, USA, for helpful discussions and critical review during the preparation of the manuscript.